lds

Dijkstra's arguments [1] about the benefits of 0-indexing are about tricky edge cases when coding with arrays. (The example you give is not a correctness problem, but a memory-efficiency issue: if width=1, you'll not be using at least half of the structure's indices that you allocated). But Lua doesn't have arrays as a type, it only has a way of iterating over the array part of tables, its map data structure. Idiomatic Lua code for tables doesn't run into the kind of edge cases Dijkstra presents, since the most readable code uses iterators to yield the values of array indices when one wants to access the array part of tables. It makes perfect sense to add arrays as a userdata type. For these, there are benefits of 0-indexing, as seen in https://github.com/neomantra/lds - although mixing 0-indexed [1]: https://www.cs.utexas.edu/users/EWD/t

I hear ya.... try: * Building with LJ_GC64 mode. [1] It is newer but people are using it and it has been merged by Mike Pall into v2.1. Still beta though. * Using FFI to allocate large objects or large pools of small objects. I made some simple containers to help with the FFI parts [2] and also a (not recently updated) jemalloc binding [3] to help tune memory usage. Here's a gist I made with some experimental results [4]. [1] https://github.com/LuaJIT/LuaJIT/issues/25 [2] https://github.com/neomantra/lds [3] https://github.com/neomantra/luajit-jemalloc [4] https://gist.github.com/neomantra/9122165